The sprinters are crouched in the starting blocks of the Generative AI race. Where will the track take humanity – is unidentified. The destination – undisclosed. Next is an explosive horizontal burst in every direction. Discussions to explore the risks of misuse of frontier generative AI, at Bletchley Park can only address the unusual challenges that these risks pose, including cyber and biological attacks. But AI is Janus faced.

Bard, Pi, Claude, Replika, Jasper, and other generative AI models are levelling the playing field. ChatSonic can generate blogs, emails, tweets, product descriptions, and ads. Some AI models are designed specifically for programmers. Others offer a supportive and smart AI design that can mimic a therapist-like approach to mental health.

Many more support multi-turn dialogues, some predict trends in fashion for manufacturers, others can create a complete advertising campaign, or original designs for a complete Carnival Band, while others support text and image creation.

One side-effect of synthetic creative outputs is that society may begin to appraise the creativity of human actors through a different lens. Bard is powered by Google’s PaLM 2 LLM and offers a unique experience to conversations. Replika focuses on companionship and deep relationships.

The “creator economy” is valued at around US$14bn per year. New digital channels have empowered writers, podcasters, painters, and musicians to connect with practitioners and collectors directly.

Substack, Flipboard, and Steemit enable creatives to become brand managers of their work. Generative AI applications like Midjourney are changing the status of how creative people work using data lakes and user feedback.

Vivienne Tam, a celebrated Hong Kong fashion aficionado has embraced the use of “Aida”. Aida is short for AI-based Interactive Design Assistant for Fashion. Aida was created in the AiDLab, a research centre that is a collaboration between the Royal College of Arts and the Hong Kong Polytechnic University. In under a few seconds, Aida can generate outfits by mining sketches, street fashion trends, fabrics, trims, patterns, lookbooks, and in-vogue colour palettes.

What does this mean for the unprotected Carnival designs of Harold Saldenah who was born in 1925 in Belmont, East Port of Spain, and George Bailey who while attending Tranquillity, studied drawing, painting, and sculpture under the tutelage of M.P. Alladin? This has raised the spectre of intellectual property infringements created by the unregulated and unlimited use of generative AI in costume and fashion design.

Amazon has decided to enter the race by partnering with Anthropic. Anthropic will use Amazon’s Trainium and Inferentia chips which are specifically designed for generative AI applications to design new AI models. This alliance parallels Microsoft’s partnership with Open AI the creator of ChatGPT. The Amazon Anthropic partnership signals a move to rival Nvidia’s dominance in the lucrative and talent thin market for specialist AI chips. Anthropic brings “Claude” into the race. Claude is an open rival to OpenAI’s ChatGPT.

Amazon is a digital shelf behemoth and a significant provider of cloud-computing services. Simply put – Amazon rents out computing power to clients everywhere on the planet. This capability is housed in humongous warehouses stacked with computers called data centres that provide other companies with data storage and processing services.

In search of Real Intelligence, Anthropic had Claude 2 take the GRE and also tested it on a range of standard benchmarks used to test AI systems. To be able to compare the two models OpenAI used many of the same benchmarks on GPT-4. Claude 2 placed in the 95th, 42nd, and 91st percentile on the verbal, computational, and writing test batteries respectively. GPT-4 had scores in the 99th, 80th, and 54th percentile. Differences in testing conditions make it difficult to draw conclusions, especially since Claude 2 was given a chain-of-thought prompt. Moreover, Claude 2 was provided with examples of GRE questions whereas GPT-4 was not.

Claude 2 is an enhancement of Anthropic’s Claude 1.3, mostly in terms of its capacity to write code using written instructions and the size of its “context window,” which now permits a user to input entire books and interrogate Claude 2 on their content. These improvements place Claude 2 in the same league as GPT- 4.

However, they both still show stereotypical bias and can hallucinate or make things up. The Governance of AI remains worrisome. Commercial incentives or national security directives could propel competitive dynamics between AI labs or among nations, leading to misadventures.

The field is getting cluttered quickly as other cloud providers create new products and are investing more in generative AI. Amazon developers will now be able to use Claude2 to create novel applications for new customers, enhance the experience of existing ones, and intertwine AI-powered upgrades.

The field of competitors is not lacklustre. Rather it is full of talent. Microsoft operates Azure. Azure is a cloud computing enterprise and has an arrangement with OpenAI. The partnership of ChatGPT and Microsoft is creating an abundance of unbelievable advances including an intelligent assistant for Microsoft 365, and Copilot.

Amazon will invest US$4bn in Anthropic which is a San Francisco-based AI firm. Anthropic is a 2021 AI safety and R&D startup that is now competing with Google DeepMind and OpenAI.