While introducing a world-leading industry standard that requires tech companies to eradicate such content from AI-powered search engines, Australia’s eSafety Commissioner warned that artificial intelligence tools might be used to generate images of child abuse and terrorist propaganda. The warning came in conjunction with the announcement of the standard.
Big tech companies like Google, Microsoft’s Bing, and DuckDuckGo are required by the new industry code governing search engines, which will be detailed on Friday, to remove content related to child abuse from their search results and to take steps to prevent generative AI products from being used to generate deepfake versions of that content. In addition, the code mandates that these companies take steps to ensure that generative AI products cannot be used to generate deepfake versions of that content.
The eSafety Commissioner, Julie Inman Grant, stated that the companies themselves have to be at the forefront of decreasing the dangers that can be caused by their products. She stated that they are receiving ‘synthetic’ child abuse material. “We are seeing ‘synthetic’ child abuse material come through,” “Propaganda is being created by terrorist organizations with the help of generative AI. This is already taking place. It’s not a make-believe concept at all. We were of the opinion that it should be addressed.”
Both Google and Microsoft have recently revealed their intent to include their respective artificial intelligence (AI) tools, ChatGPT and Bard, into their respective consumer-facing search engines. According to Inman Grant, the development of artificial intelligence technologies necessitated a rethinking of the “search code” that covered those platforms.
According to the eSafety Commissioner, the prior version of the code only covered content that search engines returned in response to inquiries and did not include content that these services could produce themselves. The new legislation will force search engines to continually review and enhance their artificial intelligence systems. This is to ensure that “class 1A” material, which includes material that promotes terrorism, extreme violence, and the sexual exploitation of children, is not returned in search results. This can be accomplished by delisting and blocking such search results.
In addition to this, the businesses will be obliged to conduct research on technology that would assist users in detecting and identifying deep fake images that are accessible via their services. It is believed by the eSafety Commission to be one of the first frameworks of its sort ever developed anywhere in the globe.
According to Inman Grant, the rapid growth of artificial intelligence has been likened to a “arms race.” He also stated that the progress of the technology had “caught most policymakers, regulators, and countries on the hop.”
“As a regulator and an implementor of policy, generative AI suggests to me that we need to think differently about how we use regulatory tools,” she said. “It’s time to rethink the way we use regulatory tools.” “If we want to get ahead of these issues, I believe that the regulatory scrutiny is going to need to be at the design and deployment phase.”
Instead than playing “whac-a-mole” as problems arise, according to Inman Grant, more regulation of these technologies needs to be built into the system from the beginning. She compared the situation to the law that requires automakers to include seatbelts in their vehicles.
According to Inman Grant, “it makes business sense to do this upfront, rather than see what happens with generative AI,” and he claimed this. “It’s time for the tech industry to buckle up,” said one analyst.
According to the eSafety Commissioner, regulators are aware of criminal actors who are utilizing new artificial intelligence techniques for unlawful objectives, including the generation of material involving child abuse. According to Inman Grant, the new regulations will require technology businesses to not just lessen the harm that can be done on their platforms, but also to focus on developing tools that promote improved safety, such as the ability to recognize deep false photos.
“As these tools become more accessible to the general public, sexual predators may use them to create synthetic child sexual abuse material according to their preferences, or they may use anime; the possibilities are virtually endless.” She explained, “We need to know the companies are thinking about this and installing appropriate guard rails.”
On Thursday, the Attorney General, Mark Dreyfus, informed parliament of separate work being done by the Australian Federal Police utilizing AI to detect child abuse material rather than the present manual analysis of images. This work is being done in place of the current assessment of photographs.
This program has produced a new tool that asks adults to contribute photographs of themselves when they were younger in order to assist train an artificial intelligence model. Dreyfus stated that he would be submitting a photograph of himself for consideration by the program.