Publishers and entertainment industry groups have recently expressed their concerns, imploring the government to dismiss proposed exemptions to the Digital Millennium Copyright Act. These exemptions serve to increase protections around text-and-data mining and would permit hackers to scrutinize biases in artificial intelligence models, subsequently raising issues around infringement and cybersecurity.
The proposed exemptions could considerably broaden the landscape in which text-and-data mining activities take place. Such activities are not inherently illegal; in fact, they can be instrumental in conducting analytical studies and generating insights from large sets of data. However, the provisions presented in the exemptions give rise to certain qualms, specifically surrounding infringement.
Typically, text-and-data mining involves copying large quantities of data, an act that – even though it’s performed for analysis purposes and the results are generally not shared – can infringe upon copyright laws. Therefore, while the proposed exemptions would offer a legal safeguard, they are also seen as a potential door to copyright abuse.
Similarly, allowing hackers to study biases in artificial intelligence models is seen as a double-edged sword. While it’s unquestionably crucial to promote transparency and fairness in AI, this provision could compromise cybersecurity. In studying these models, hackers would necessarily have to probe and potentially exploit their weaknesses, thereby exposing the models to potential ill-intended uses.
Overall, it’s clear these proposed exemptions are dividing opinion. More safeguards might be needed to ensure that they are not misused. For legal professionals, it’s crucial to stay abreast of the situation and understand the potential implications for their operations. For a more detailed examination of the proposal and the various responses to it, you can read the in-depth exploration by Ivan Moreno on Law360.