New regulations seek to mitigate risks from artificial intelligence (AI) systems like flawed algorithms or biased data while maintaining incentives for ongoing innovation. Getting the balance right poses challenges.
What's in the EU's new AI Act?
The act creates a graduated risk framework for AI systems used in areas like critical infrastructure, employment decisions and law enforcement. Clear transparency rules apply even for low-risk AI. Strict restrictions can halt systems causing systematic harm.
Oversight focuses on high-risk categories like:
- Biometric identification tech
- Infrastructure management
- Hiring tools
- Legal decision support
Controversially, unproven "foundational models" like chatbots escape uniform rules which troubles privacy advocates. The EU wants safeguards against abusive content or data exploitation.
Industry groups argue heavy regulations could slow breakthroughs from AI systems that remain works in progress. They prefer voluntary standards and self-governed guidelines.
Michael Bacina, Partner at Piper Alderman advising in Digital Law said,
“Much like the crypto-space, the fears around AI and regulation may be lessened as education as to how large language models work spreads. My concern is far more around scams and impersonations and how the authenticity of brands and businesses can be retained in a world where the cost of impersonating a business is dramatically lowered by LLMs. Regulation which is focused on the risks of AI may miss or pay scant attention to the very real and practical damage being caused by this problem.”
Implications for IP and Patent Law
By restricting certain types of AI tech, regulations impact intellectual property (IP) rights and technology licensing. Regulations may alter commercialisation incentives or cause inventors to alter their filing strategy to file in less regulated domains".
Trends towards mandatory transparency in machine learning algorithms may also compel tech firms to publicly disclose features prior to the publication of those features in corresponding patent applications or where patents have not been filed make replication easier.
What Should Australian Firms Do?
Navigating the complex intersection of AI innovation and regulation poses growing challenges for technology companies. IP-reliant tech companies do well to track emerging governmental frameworks regulating AI systems and to check if mandatory standards impact research pipelines. It is also possible to engage in public consultations to bring industry perspectives to the innovation vs risk balance.
AI oversight across regions remains fluid with Australia studying moves by Europe and the quest for ethical yet innovative AI persists. During this period of uncertainty, Baxter IP will continue to optimise IP protection in light of latest developments. Contact our IP Attorneys today to explore your options in this rapidly-evolving domain.