AUSTRALIA Trends and Developments Contributed by: John Lee, Michael Williams, Siabon Seet and Vanessa Farago-Diener, Gilbert + Tobin
On the other side of the coin, a question arises as to whether AI outputs may constitute trade secrets or confidential information of the user (or their employer). Again, much will depend on the circumstances of the case, including the tool in question, and any contractual arrange - ments or terms of use governing the relation - ship between provider and user (or employer). The more control an organisation has over the inputs and the outputs (including restrictions on access, sequestration of data, etc), the better able an organisation may be to argue that AI out - puts “belong” to it and constitute its confidential information. Australia is yet to enact any legislation specifical - ly regulating AI technologies, but the regulation of AI is very firmly on the Australian government’s radar. On 5 September 2024, the government released a proposals paper for Introducing man - datory guardrails for AI in high-risk settings (the Proposals Paper), which set out various options being considered in order to mandate guardrails on those developing and deploying AI in Aus - tralia in high-risk settings. In parallel, the federal Attorney-General circulated a discussion paper to the recently-established Copyright and Arti - ficial Intelligence Reference Group (CAIRG), to seek its views “on whether there were any spe- cific copyright-related AI transparency issues on which the government should consider urgent action through AGD’s work on copyright law and AI (including because they may fall outside of the scope of the transparency proposals in the Pro- posals Paper)” . The discussion paper, Copyright and AI Transparency Issues, sought views from CAIRG members on transparency issues relating to AI inputs and AI outputs. Relevantly, CAIRG respondents generally supported transparency requirements relating to inputs for AI develop - ment if they “apply at a high level (eg, a summary, rather than in detail) and incorporate protections
or carve-outs for sensitive commercial informa- tion (such as trade secrets).” In addition, in November 2024, the Senate Select Committee on Adopting Artificial Intel - ligence published a report on the opportuni - ties and impacts arising from the uptake of AI technologies in Australia. The Committee’s core recommendation was that the government intro - duce new, whole-of-economy, dedicated legisla - tion to regulate high-risk uses of AI, adopting a principles-based approach to defining high-risk AI uses, supplemented by a non-exhaustive list of explicitly defined high-risk AI uses (with the non-exhaustive list explicitly including general- purpose AI models such as large language mod - els). When and if such legislation is enacted, it is unclear to what extent it will address IP and confidentiality issues arising in connection with the use of AI technology. Strategies for preventing disclosure of trade secrets in the AI age The ability of AI to analyse data, undertake pat - tern recognition and process data at a scale and speed impossible for humans can give rise to competitors using AI tools to obtain a competi - tive advantage and can even lead to the reverse engineering of products and processes. To protect their trade secrets and confidential information, companies can implement a num - ber of measures specifically directed at the use and implementation of AI tools. • Internal trade secret protection policies – companies should have in place robust poli - cies to protect their trade secrets and confi - dential information. Policies should include guidance as to what information can be entered into AI tools, with appropriate qualifi -
37
CHAMBERS.COM
Powered by FlippingBook