A group of tech heavy hitters is joining forces to boost artificial intelligence with one key goal: a coordinated effort to enhance safety amid the rising fears of AI risk, including misinformation, cyber threats, discrimination and data theft.
The recently launched AI Alliance, started by IBM and Facebook owner Meta, brings together government, academia and industry to cooperate on standards, safety, and security. Other members of the new global group include AMD, Intel, Dell and Sony, plus federal agencies such as NASA and universities such as Yale.
Expect promising work from the group as it promotes open tech standards, which make it easier to persuade software developers to build new apps and tools. The group will also create a catalogue of vetted security, safety and trust tools with one unstated mission: Shaping government regulation to avoid strict laws. The group will also host events showing how members use AI responsibly and safely.
Tech companies lobbying to avoid strict AI regulations didn’t work in Europe. The European Union is moving ahead with its AI Act, which is set to take effect in 2025. The regulations will involve high levels of transparency and reporting, a headache for many companies.
They’ll also ban a slew of AI applications, such as biometrics, that help categorise people by sensitive characteristics (religion, politics, race, etc.) and emotion recognition in the workplace or at school. Some bans are very broad, such as barring “AI systems that manipulate human behaviour to circumvent their free will.” It will take much more commentary from regulators to understand how the rules will work in practice as hefty fines are coming for companies found in violation.
The U.S. and other countries will be watching closely. Any legislation from Congress will likely be much less strict than the EU’s, giving the U.S. an ongoing edge. Alphabet’s new product exemplifies today’s AI arms race. The firm’s twist: Its new Gemini model can “understand” a range of inputs from text, computer code, audio, image and video, so the AI chatbot can answer questions, interpret data and more. Google has three versions, including an energy-saving one made for mobile devices.
Some notable AI firms and tools include Microsoft-backed OpenAI with its ChatGPT, Stability AI and its image creator Stable Diffusion, Cohere, a business-focused AI firm, Hugging Face, an online hub for start-ups to find AI models and Anthropic’s Claude, a popular AI assistant. Plus, big tech companies such as Amazon, Meta and Apple.
Meanwhile, how is emerging AI going to help manufacturers? 40% of manufacturing executives in a recent survey said to start by analysing data to help evaluate and organise plant floors. That’s especially true for plants with web-connected machinery. For example, Nvidia’s Omniverse software can use AI to model logistics and robots at a plant before production starts. Virtual reality can also help train employees, and executives think AI will help with predictive maintenance, process improvements, and quality control.
However, AI tech can have large up-front costs, a looming challenge for small manufacturers, and the testing phase is lengthy. Small manufacturers may lack enough data to train the AI system well enough. However, the software’s costs will continue to fall, and more digitised data will become available, making the tech more accessible.
