1. Introduction

Autonomous Evidence Intelligence Governance in Thailand concerns the legal regulation of artificial intelligence (AI), automated evidence-analysis systems, digital forensic technologies, machine-learning platforms and intelligent judicial-support systems used to collect, authenticate, classify, preserve, analyse and present evidence in civil proceedings.

The concept involves the integration of autonomous technologies into the Thai civil justice system while maintaining:

Authenticity and reliability of evidence.

Procedural fairness and equality between litigating parties.

Privacy and personal-data protection.

Human supervision over automated evidentiary decisions.

Accountability for manipulated, fabricated or AI-generated evidence.

Transparency in algorithmic evidence assessment.

Judicial independence and the right to challenge evidence.

Thailand does not currently have a separate comprehensive statute governing autonomous evidence intelligence in civil litigation. Instead, the applicable framework is derived from existing legislation, particularly the Civil Procedure Code, Electronic Transactions Act B.E. 2544 (2001), Computer Crime Act B.E. 2550 (2007), Personal Data Protection Act B.E. 2562 (2019), and Civil and Commercial Code.

An important distinction is that electronic evidence is legally admissible, but AI-generated evidence is not automatically authentic or reliable merely because a computer system produced it.

2. Legal framework governing autonomous evidence intelligence

A. Electronic Transactions Act B.E. 2544 (2001)

Section 11 is fundamental to electronic evidence governance.

It establishes that electronic data cannot be rejected in civil, criminal or other legal proceedings solely because it exists in electronic form.

In assessing evidentiary weight, courts consider:

The reliability of the method used to generate the information.

The reliability of electronic storage and communication.

The integrity and completeness of the information.

The method used to identify the originator.

Other relevant circumstances.

This provision is particularly important for:

AI-generated forensic reports.

Electronic contracts.

Automated transaction histories.

Digital signatures.

Blockchain records.

AI-assisted document authentication.

The provision establishes admissibility, not conclusive proof.

:+1

B. Civil Procedure Code

The Thai Civil Procedure Code regulates the presentation, production, examination and assessment of documentary and other evidence.

Section 93 – Original document principle

The original-document principle is relevant when parties rely on electronic documents, copies, screenshots or exported records.

The court must consider whether the applicable statutory exceptions permit a copy or alternative evidence.

Section 125 – Objections to documentary evidence

Parties must raise appropriate objections to documentary evidence in accordance with procedural requirements.

This is important where one party relies on AI-generated reports, electronic communications or machine-produced records.

Sections 84 and related evidentiary provisions

The court determines the relevance, admissibility and weight of evidence under the applicable procedural rules.

An AI system cannot independently replace the court's responsibility to evaluate disputed facts.

ThaiLawOnline+1

C. Computer Crime Act B.E. 2550 (2007)

This legislation is relevant when autonomous evidence systems obtain information from computers, servers, cloud infrastructure or communication networks.

Important provisions include:

Section 18: Provides investigative powers for obtaining computer data and relevant information, subject to statutory requirements.

Section 25: Regulates the evidentiary use of computer data obtained under the Act and excludes evidence obtained through prohibited inducement, deception or other unlawful means.

Section 26: Establishes computer-traffic-data retention obligations for service providers.

These provisions are relevant to digital forensic investigations, cyber-fraud litigation and the preservation of evidence.

:+2

D. Personal Data Protection Act B.E. 2562 (2019)

Autonomous evidence intelligence may process:

Personal identification records.

Biometric information.

Financial information.

Health records.

Location histories.

Private communications.

Behavioural profiles.

The PDPA requires lawful processing and appropriate safeguards.

Particularly relevant provisions include:

Section 19: Consent requirements and exceptions.

Section 23: Transparency and privacy-notice obligations.

Section 24: Lawful bases for processing personal data.

Section 26: Additional protection for sensitive personal data.

Section 37: Data-controller security and compliance responsibilities.

Sections 77–78: Civil liability and compensation.

An evidence-intelligence operator may therefore face civil liability when personal data is improperly collected, disclosed or processed.

E. Civil and Commercial Code

The Civil and Commercial Code provides the foundation for compensation and civil responsibility.

Section 420: A person who wilfully or negligently unlawfully injures another person's life, body, health, liberty, property or rights may be liable in tort.

Sections 421–422: Address abusive exercise of rights and certain presumptions concerning wrongful acts.

Section 425: Relevant to employer responsibility for wrongful acts committed by employees in the course of employment.

Sections 438 and 442: Relevant to judicial assessment of compensation and contributory fault.

These provisions can apply where defective autonomous evidence systems cause financial loss, reputational damage or infringement of legally protected rights.

3. Autonomous evidence intelligence: principal governance mechanisms

A. Automated evidence collection

AI-powered systems can collect evidence from:

Email communications.

LINE and other messaging applications.

Digital banking records.

Cloud-storage systems.

Business transaction databases.

CCTV and surveillance systems.

Smart devices and IoT networks.

The legal issue is whether the collection method was lawful and whether the resulting evidence accurately represents the original information.

For example, a company using AI to investigate employee fraud must distinguish between lawfully obtained company records and private information obtained without an adequate legal basis.

B. Intelligent evidence authentication

Evidence-intelligence systems can use machine learning, cryptographic hashes, digital signatures and metadata analysis to assess whether a document has been altered.

An appropriate authentication process should establish:

Original source of the record.

Date and time of creation.

Identity of the relevant user or system.

Record of subsequent modifications.

Integrity of storage.

Reliability of the authentication technology.

A cryptographic hash may help demonstrate that a file has not changed since a particular point in time. However, it does not independently establish that the original file was truthful.

C. AI-generated evidence analysis

AI can analyse large collections of documents and identify:

Inconsistent statements.

Suspicious financial transfers.

Contradictory contractual provisions.

Unusual transaction patterns.

Possible document forgery.

Relationships between individuals and organisations.

Such analysis may assist lawyers, forensic experts and judges, but the underlying source material and analytical methodology must remain open to appropriate examination.

D. Autonomous evidence preservation

Evidence preservation is especially important because digital evidence can be altered, deleted or overwritten.

A sound governance system should include:

Secure backups.

Timestamped records.

Chain-of-custody documentation.

Access-control mechanisms.

Tamper-evident audit logs.

Preservation of original electronic files.

Documented forensic extraction procedures.

The Computer Crime Act's Section 26 retention requirements are particularly relevant to computer-traffic data. The statutory retention period should not be confused with an unlimited obligation to preserve every category of digital evidence.

:+1

E. Algorithmic transparency

Where AI produces an evidentiary assessment, the opposing party should have a meaningful opportunity to challenge:

The source data.

The reliability of the algorithm.

The assumptions used.

The possibility of biased training data.

The probability of false positives.

The expert's interpretation.

The relationship between the AI output and the disputed fact.

An AI-generated probability score cannot automatically become proof of civil liability.

F. Human judicial supervision

The final determination of disputed facts remains a judicial responsibility.

AI may assist with document classification, chronology construction, evidence comparison and transcription, but should not independently determine credibility, liability or the outcome of a civil claim.

4. Important Thai Supreme Court case laws

The following decisions provide relevant judicial principles for autonomous evidence intelligence. Some concern electronic evidence directly, while others concern documentary reliability, hearsay, civil proof and procedural treatment of evidence. They should not be misrepresented as decisions specifically regulating modern autonomous AI systems.

Case 1: Supreme Court Decision No. 3046/2537

Subject: Recognition of electronic evidence before the Electronic Transactions Act

This decision is frequently cited in discussions of the historical difficulties surrounding electronic evidence in Thailand.

The pre-electronic-transactions legal environment was less accommodating to electronic records as substitutes for conventional documentary evidence.

Legal significance:

Demonstrates the historical importance of documentary formalities.

Illustrates the development towards technology-neutral evidentiary rules.

Provides historical context for Section 11 of the Electronic Transactions Act.

Supports the need for clear standards governing digital-document authenticity.

Application: Autonomous evidence systems should preserve legally recognisable records and ensure that electronically generated information can be examined in court.

Case 2: Supreme Court Decision No. 7264/2542

Subject: Reliability of computer-generated banking records

This is an important Thai decision concerning computer-generated information.

The dispute involved missing bank funds and discrepancies between a customer's passbook and the bank's computer records.

The Supreme Court considered the bank's established computer technology, the quality of its systems and the competence of its personnel. It accepted the reliability of the computer records in the circumstances.

Legal principles:

Computer-generated records can provide persuasive evidence.

Reliability depends on the system's technical and operational characteristics.

The competence of system administrators is relevant.

Computer records must be assessed alongside surrounding circumstances.

Digital records may help establish the occurrence of financial wrongdoing.

Application: AI-based banking surveillance and automated financial investigations may benefit from this reasoning, provided their underlying systems and records can be independently verified.

Case 3: Supreme Court Decision No. 6010/2564 (2021)

Subject: Independent assessment of evidence in civil proceedings

This decision concerned alleged fraudulent transfers of money into different bank accounts.

The Supreme Court explained that where the criminal proceedings against a defendant had ended through withdrawal of the prosecution without a final criminal judgment, the civil court was not bound by findings that had never been finally determined.

The civil court had to assess the evidence presented in the civil proceedings independently.

The decision also addressed the responsibility of individuals who knowingly participated in the fraudulent receipt of money.

Legal principles:

Civil liability requires independent evidentiary assessment.

A withdrawn criminal prosecution is not equivalent to a final criminal judgment.

Financial transactions must be assessed in their factual context.

Knowledge and participation must be established through sufficient evidence.

Application: An autonomous intelligence platform must not automatically classify someone as civilly liable merely because a related criminal investigation or prosecution exists.

Case 4: Supreme Court Decision No. 3298/2564 (2021)

Subject: Hearsay evidence and assessment of reliability

This decision concerns the treatment of hearsay evidence under Section 95/1 of the applicable procedural framework.

The Supreme Court recognised that hearsay is not invariably excluded. Evidence may be admitted under the statutory exceptions where its source, circumstances and nature make it sufficiently credible.

The decision illustrates the importance of evaluating the reliability of information rather than applying an inflexible rule.

Legal principles:

Hearsay must be assessed under the applicable statutory exceptions.

Source reliability is relevant.

Surrounding circumstances affect evidentiary weight.

Supporting evidence may strengthen the reliability of disputed information.

Application: AI systems that summarise third-party communications must identify whether the information is firsthand, hearsay or independently corroborated.

Case 5: Supreme Court Decision No. 5954/2567 (2024)

Subject: LINE communications and inadmissible hearsay

This decision is particularly relevant to digital-evidence governance.

The defendant relied on LINE communications exchanged between two other individuals who were not called as witnesses.

The Supreme Court treated the statements as hearsay under Section 95/1 and found that the applicable exception had not been established.

Legal principles:

Electronic communication does not automatically overcome hearsay restrictions.

A screenshot is not necessarily sufficient proof of the truth of its contents.

The identity and availability of the original speakers matter.

The circumstances surrounding the communication must be examined.

Electronic records remain subject to ordinary evidentiary principles.

Application: AI systems that extract statements from LINE, Messenger or other applications must distinguish between the existence of a message and the truth of the statement contained in it.

Case 6: Supreme Court Decision No. 5661/2567 (2024)

Subject: Documentary evidence and procedural objections

This decision concerned documents relating to a road construction permit and the legal status of a disputed road.

The Supreme Court considered the interaction between Sections 93(4) and 125 of the Civil Procedure Code.

The decision illustrates that a party's failure to make an appropriate objection to documentary evidence during the proceedings can affect whether that evidence is admitted.

Legal principles:

Documentary admissibility is affected by procedural requirements.

Objections must be raised at the appropriate stage.

Copies may be accepted in circumstances authorised by law.

Procedural conduct influences the treatment of documentary evidence.

Application: Autonomous litigation platforms should track deadlines for evidentiary objections and flag procedural risks rather than treating all uploaded documents as automatically admissible.

Case 7: Supreme Court Decision No. 1112/2566 (2023)

Subject: Electronic LINE messages and proof of a loan

The case concerned a claimed loan of 7.8 million baht supported by LINE communications.

The Supreme Court considered whether the electronic messages established a legally sufficient written acknowledgement of the alleged loan.

Although electronic messages could qualify as electronic data and satisfy relevant requirements concerning electronic writing and signatures, their actual contents did not sufficiently establish the alleged loan.

Legal principles:

Electronic communications can have legal evidentiary significance.

Electronic signatures and written-form requirements must be examined.

The substance of the messages matters.

A transfer of money does not independently prove that the transfer was a loan.

Context and corroboration are important.

Application: AI contract-analysis systems must distinguish between evidence of a payment and evidence of the legal purpose of that payment.

Case 8: Supreme Court Decision No. 3622/2568 (2025)

Subject: LINE messages as electronic written evidence of a loan

This decision provides an especially useful illustration of the difference between electronic admissibility and substantive evidentiary sufficiency.

The dispute involved two separate transfers.

For the first transfer of 200,000 baht, the LINE conversation expressly identified the transaction as a loan, and the borrower responded in a manner supporting that understanding.

The second transfer of 100,000 baht was not supported by sufficiently clear written evidence establishing the same loan arrangement.

The Supreme Court distinguished the two transactions despite their appearance in the same conversation.

Legal principles:

Electronic communications can satisfy written-evidence requirements.

Context and meaning determine the legal significance of messages.

An electronic signature does not automatically establish every contractual element.

Each disputed transaction requires individual examination.

AI interpretation of messages must preserve their complete conversational context.

Application: Automated evidence-intelligence platforms should not infer an entire contractual relationship from isolated messages or payment records.

The decisions discussed above illustrate a progression from historical uncertainty about electronic records to a more sophisticated approach involving authenticity, hearsay, procedural compliance and the substantive meaning of digital communications.

silklegal.com+6

5. Civil Court Regulations on AI Technology B.E. 2568 (2025)

An important development in Thailand is the introduction of the Civil Court Regulations on the Use of AI Technology in the Preparation of Pleadings or Other Documents for Submission to the Court B.E. 2568 (2025).

These regulations are particularly relevant to autonomous evidence intelligence because AI can generate or modify material included in court submissions.

Their principal requirements include:

A. Disclosure of AI-generated material

Content generated using AI must be appropriately identified in court documents.

B. Verification by lawyers and parties

The submitting party and lawyer remain responsible for checking the accuracy of AI-generated facts and legal provisions.

C. Accountability

A party cannot avoid responsibility merely by claiming that an AI system generated inaccurate information.

D. Procedural consequences

Non-compliance may result in sanctions and may constitute contempt of court under the relevant procedural provisions.

These requirements establish an important principle: AI can assist in preparing litigation documents, but responsibility remains with the human submitting them.

Oxford Institute of Technology and Justice

6. Civil liability arising from autonomous evidence intelligence

Autonomous evidence systems can create different categories of civil liability.

A. Liability of AI developers

Developers may face liability where negligent software design, inadequate testing or failure to address foreseeable risks causes legally recognised harm.

Examples include:

Incorrect document authentication.

Failure to identify manipulated records.

Defective forensic classification.

Inaccurate financial evidence analysis.

B. Liability of evidence-platform operators

Operators may be responsible where they negligently configure systems, improperly preserve records or fail to maintain adequate security.

C. Liability of professional users

Lawyers, forensic experts and organisations using autonomous evidence platforms must exercise appropriate professional care.

Reliance on automated outputs without reasonable verification may create liability depending on the applicable duties and circumstances.

D. Liability of organisations

Organisations using AI to investigate employees, customers or business partners must ensure lawful collection, processing, retention and disclosure of personal information.

7. Major challenges in Thailand

1. Deepfake evidence

AI can generate synthetic audio, photographs, videos and documents.

The main difficulty is distinguishing authentic recordings from manipulated material.

Thai Criminal Code Section 180 is relevant to the production or presentation of false evidence in judicial proceedings. Its application must be distinguished from the separate civil question of whether a litigant has proved a claim.

2. Algorithmic bias

An AI system may incorrectly associate individuals with fraudulent activity because of biased or incomplete training data.

This can result in wrongful allegations and financial or reputational harm.

3. Evidence-chain manipulation

Digital records may be altered during extraction, transmission, storage or presentation.

A reliable chain of custody is therefore essential.

4. Privacy and surveillance

Autonomous evidence systems may process sensitive information without adequate legal justification.

The PDPA provides an important framework for protecting individuals.

5. Explainability

An AI platform may identify suspicious transactions without explaining the underlying reasoning.

Courts and opposing parties need sufficient information to examine the reliability of the result.

6. Cross-border evidence

Cloud-based evidence may be stored outside Thailand.

This creates challenges involving jurisdiction, data access, preservation, privacy and international cooperation.

7. Judicial dependence on automation

Overreliance on automated recommendations may undermine independent judicial evaluation.

Human judicial supervision must remain central.

8. Proposed autonomous evidence intelligence governance model for Thailand

Thailand could develop a comprehensive framework containing the following components.

Proposed governance architecture

Level 1 – Evidence collection

Lawful acquisition, source identification and consent or other legal basis

Level 2 – Evidence authentication

Digital signatures, metadata, hashes and forensic verification

Level 3 – AI evidence intelligence

Classification, anomaly detection and cross-document analysis

Level 4 – Independent human verification

Expert examination, disclosure, objections and cross-examination

Level 5 – Judicial determination

Independent judicial assessment of admissibility, credibility and evidentiary weight

Cross-cutting safeguards

PDPA compliance and privacy protection.

Cybersecurity and evidence preservation.

Algorithmic transparency and auditability.

Human accountability.

Effective remedies for affected parties.

9. Recommended legislative and institutional reforms

Thailand could strengthen this field through the following reforms:

1. Autonomous Evidence Intelligence Act: Establish specific rules for AI-assisted collection, analysis and presentation of evidence.

2. Digital Evidence Authentication Standards: Establish uniform technical requirements for metadata, timestamps, digital signatures and forensic integrity.

3. AI Evidence Disclosure Rules: Require disclosure of material AI involvement in the preparation or analysis of evidence.

4. Independent Algorithmic Audits: Introduce audits for high-impact evidence-intelligence systems.

5. Deepfake Detection Standards: Establish procedures for examining synthetic audio, video and photographic evidence.

6. Expert Witness Accreditation: Develop appropriate standards for digital forensic and AI evidence experts.

7. Judicial Technology Training: Strengthen judicial understanding of algorithmic errors, data manipulation and electronic evidence.

8. Evidence Preservation Infrastructure: Develop secure systems for maintaining original records and auditable chains of custody.

9. Cross-Border Cooperation: Establish appropriate procedures for obtaining and authenticating foreign-hosted electronic evidence.

10. Civil Remedies: Clarify compensation and accountability for losses caused by negligent or unlawful autonomous evidence processing.

10. Comparative legal perspective

Thailand can combine technology-neutral electronic-evidence rules with stronger AI transparency, expert verification and judicial oversight.

11. Conclusion

Civil Law and Autonomous Evidence Intelligence Governance in Thailand represents an important emerging area of civil justice and technological regulation.

The Electronic Transactions Act establishes the admissibility of electronic information, while the Civil Procedure Code determines how evidence is presented, challenged and assessed. The PDPA protects personal information, and the Civil and Commercial Code provides remedies for wrongful conduct.

The Supreme Court decisions discussed above demonstrate several important principles:

Computer-generated records can be reliable evidence.

Electronic communications can satisfy legal requirements in appropriate circumstances.

Hearsay restrictions continue to apply to digital messages.

Electronic evidence must be evaluated according to its authenticity and context.

AI-generated analysis cannot replace independent judicial assessment.

Parties and lawyers remain accountable for material submitted to courts.

The central legal principle is that autonomous evidence intelligence should strengthen the accuracy, integrity and efficiency of Thai civil justice without replacing human responsibility, procedural fairness or judicial independence.

Research qualification: The eight decisions above are relevant judicial authorities, not eight precedents specifically deciding autonomous AI evidence governance. The 3046/2537 decision is included for historical context, and several other decisions concern criminal or general evidentiary principles rather than AI technology itself. Their application to autonomous systems is therefore analytical, not a claim that the Supreme Court has already established a comprehensive AI-evidence doctrine.