Plus, AI-generated fake news swamps TikTok
A team of scientists at the University of Washington is using AI to design synthetic proteins, accelerating the process of drug discovery and development. AI algorithms, similar to those used in image generation, are being employed to create proteins with specific shapes. This approach eliminates the need for tedious trial and error testing, enabling researchers to generate functional proteins in a matter of months. AI’s potential to revolutionize the field of science is being recognized, with experts envisioning collaborative partnerships between AI systems and human scientists in experimental planning and execution. However, challenges exist, such as the need for extensive research databases and organized data in fields like materials science. Nevertheless, the impact of AI on scientific discovery could be immense, from generating new hypotheses to continuously analyzing and updating research findings. By harnessing the power of AI, scientists are optimistic about the transformative possibilities in drug development and other scientific domains.
The U.S. Department of Commerce is considering expanding export controls on certain AI technologies, specifically frontier models, which are advanced AI programs that could have unpredictable and potentially dangerous capabilities. The controls are part of an ongoing effort to curb China’s development of AI and enhance national security. However, experts warn that such measures could lead to friction with China and hamper AI innovation in the U.S. The proposed controls target future AI models rather than physical parts, and there are concerns that broader definitions could hinder open-source collaboration and negatively impact smaller developers. The current discussions surrounding AI regulation are also influenced by increasing worries about China and are marked by confusion and a desire among policymakers to take action. While some tech companies support a licensing process for frontier AI, critics argue that these regulations distract from present-day concerns, such as privacy and copyright issues, and could promote fear-mongering against China.
A growing number of TikTok accounts are spreading misinformation by using AI-generated voices. Users are creating convincing fake audio, imitating the voices of celebrities and public figures, including former President Barack Obama. This AI voice-cloning technology has become a powerful tool in the dissemination of false information and conspiracy theories. Companies like ElevenLabs have released advanced tools that allow users to create AI puppets with just a few clicks. The proliferation of AI-generated audio poses a threat to public opinion and the upcoming 2024 election. While platforms like TikTok require labels for AI-generated content, many of the flagged videos did not have these disclosures. Social media platforms are scrambling to address this issue and have started removing accounts for posing as news organizations and spreading harmful misinformation. Detection of AI-generated audio remains a challenge, as current methods can be easily evaded. Tech companies must continuously improve detection methods to keep up with the advances in AI generation.
Geoffrey Hinton, a pioneer in AI, warns that if humans are not cautious, AI-enhanced machines could surpass human control within the next five years. Hinton raises concerns about the possibility of AI systems developing their own computer code, which could make them difficult to manage. Despite his warning, Hinton acknowledges that the full understanding of AI technology still eludes even its creators. He calls for increased research, government regulations, and international bans on AI-powered military robots. Other AI experts have differing views, arguing that humans have the ability to intervene and halt any dangerous technological advancements. As various industries already benefit from AI, Hinton emphasizes the need for timely implementation of AI regulations to ensure a balance between innovation and safety. Overall, Hinton believes that humanity is facing a critical turning point and urges tech and government leaders to make informed decisions regarding AI’s future.
A recent study published in Nature Chemical Biology reveals how machine learning is transforming the discovery of antibiotics. Researchers have successfully used AI to identify abaucin, a powerful antibiotic that targets the bacterial pathogen Acinetobacter baumannii. This breakthrough brings hope for more efficient and accelerated drug discovery processes, as traditional methods often relied on time-consuming and costly screenings of large compound libraries. By harnessing the capabilities of AI, scientists can predict the antimicrobial activity of molecules against specific pathogens, paving the way for the development of novel drugs with enhanced properties. The discovery of abaucin through machine learning represents a significant milestone in antibiotic research and underscores the potential of AI in revolutionizing the healthcare industry.
A team of researchers at Stevens Institute of Technology has developed an AI diagnostic tool to detect and classify valvular heart diseases with impressive accuracy. The tool uses a short burst of audio data captured by a contact microphone, which is then analyzed by a neural network. In just a few seconds, the AI tool can identify valvular heart disease with high sensitivity and specificity, significantly reducing the number of undiagnosed cases. The current standard stethoscope examinations only detect 44% of these cases, resulting in delayed treatment and increased costs for the healthcare system. The researchers’ innovation lies in their ability to detect multiple diseases simultaneously and accurately classify the severity of the condition. Their goal is to refine the tool further and extend its use to detect other circulatory diseases. Ultimately, this AI tool has the potential to revolutionize cardiac disease detection and ensure early treatment for patients.