I. CONTEXT: AI-ENGINEERED PLASMA-BASED AIR PURIFICATION DEVICES

• These devices use plasma technology to remove pollutants from air.
• AI engineering could optimize:
  • Electrode configuration
  • Plasma generation patterns
  • Energy efficiency
  • Airflow for maximum purification

Key legal question:
Can an AI-engineered device like this be patented, especially when AI plays a major role in the design?

II. PATENTABILITY REQUIREMENTS

To patent an AI-engineered plasma air purifier, it must satisfy:

1. Patentable Subject Matter
   • Must be a device or process (machines and methods are patentable).
   • Pure ideas, natural phenomena, and abstract algorithms are not patentable.
2. Novelty (Newness)
   • The configuration, plasma technique, or AI-based optimization must not be previously disclosed.
3. Inventive Step / Non-Obviousness
   • Must not be obvious to someone skilled in electrical engineering or air purification.
4. Industrial Applicability / Utility
   • Must actually clean air, not just theoretically work.
5. Inventorship
   • Most jurisdictions require a human inventor. AI cannot be legally recognized as inventor.

III. KEY CASE LAWS AND PRINCIPLES

1. DABUS Cases (AI Inventorship)

These cases are foundational for AI-related inventions.

(a) UK Supreme Court – Thaler v Comptroller-General (2023)

• AI named as inventor → rejected.
• Principle: inventor must be a natural person.
• Relevance:
  • If AI designs plasma electrode arrangements, human engineer must file the patent.

(b) US – Thaler v. Vidal (Federal Circuit, 2022)

• AI cannot be inventor under US Patent Act (inventor = “individual” = human).
• Application: The AI-designed plasma purifier device must be filed under a human engineer.

(c) European Patent Office (EPO) Decisions, 2022–2023

• AI cannot be listed as inventor.
• Devices are patentable if technical contribution is provided.
• For plasma purifiers, the device itself can be patented, not the AI that designed it.

(d) Australia – Thaler v Commissioner of Patents (2021)

• Initially allowed AI as inventor, later overturned → human inventor required.
• Illustrates global trend: AI = tool, human = inventor.

2. Alice Corp v CLS Bank (US Supreme Court, 2014) – Software/Algorithm Limitation

• Holding: Pure abstract algorithms cannot be patented.
• Relevance:
  • If AI contribution is purely algorithmic optimization without technical device modification, that part is not patentable.
  • But if AI designs plasma electrode configuration or airflow paths, producing technical effect, patentability is supported.

3. KSR v Teleflex (US Supreme Court, 2007) – Obviousness

• A solution is obvious if predictable from prior art.
• Application:
  • AI-designed plasma patterns may be considered obvious if similar configurations exist.
  • Must demonstrate non-obvious optimization, e.g., unique plasma pulse modulation for enhanced air purification.

4. Amgen v Sanofi (US Supreme Court, 2023) – Enablement

• Patent must enable skilled person to reproduce the invention.
• Application:
  • The filing must include:
    • Circuit design
    • Plasma chamber configuration
    • AI optimization methodology
  • Ensures human engineers can implement the AI-designed device.

5. EPO – Comvik Approach (Technical Effect Test)

• Software is patentable if it produces technical effect.
• Application:
  • AI-designed plasma device → technical effect = air purification efficiency
  • AI algorithms alone → abstract → not patentable
• Example:
  • Algorithm controlling electrode firing timing to reduce ozone → patentable as part of device.

6. Swiss Federal Administrative Court (DABUS)

• AI cannot be inventor.
• Human can file patent based on AI output.
• Relevance:
  • Human engineer interpreting AI-generated plasma design can be inventor.
  • Protects intellectual contribution while acknowledging AI tool.

7. Indian Patent Law – Software-Related Inventions

• Under Indian Patents Act, 1970:
  • Devices with technical effect using software/AI are patentable.
  • Pure AI algorithms without technical effect → not patentable.
• Application:
  • Plasma purifier with AI-optimized electrode arrangement → patentable if human inventor is listed.

IV. APPLICATION TO AI-ENGINEERED PLASMA PURIFIERS

Scenario A: Fully Autonomous AI Design

❌ Not patentable (inventorship issue)

Scenario B: AI-Assisted Design, Human Inventor

✅ Patentable if:

• AI optimizes electrode layout, voltage modulation, airflow, etc.
• Human interprets, validates, and files patent
• Device produces technical effect (actual air purification)
• Enablement provided

Scenario C: Pure Algorithm / Software

❌ Patentable? Likely rejected under Alice / Comvik:

• Only abstracts like “AI optimizes air” without technical implementation are insufficient

V. STRATEGIC PRINCIPLES

1. Human Inventorship is Mandatory
2. Technical Effect Required (air cleaning, energy efficiency, reduced ozone)
3. Enablement is Essential (allow reproduction)
4. Obviousness Standard Must Be Met (AI outputs can reduce inventive step)
5. AI = Tool, Not Inventor

VI. CONCLUSION

For AI-engineered plasma-based air purification devices:

• Patentable if:
  • Human is inventor
  • Device design solves a real technical problem
  • AI is used as a tool, not inventor
  • Device is sufficiently disclosed and enabled
• Not patentable if:
  • AI is listed as inventor
  • Design is purely abstract algorithm
  • AI output is obvious from prior art