IBM Meta Dell Safe AI Innovation artificial intelligence

AI Alliance: IBM, Meta, Dell and 50+ Founding Partners Pursue Open, Transparent and Safe AI Innovation

Sorting valuable insights from the hype in the self-promotional high-tech industry is no easy task. That is especially true when valuable or potentially transformative technologies become increasingly market-worthy. As an example, one needs to look no further than the melodrama surrounding Generative Artificial Intelligence (GenAI) and high-profile players, like OpenAI and its Chat GPT service during the past few months.

However, while media outlets have obsessed over OpenAI’s’s adventures, including massive investments by Microsoft, the tumultuous ousting of its CEO, Sam Altman, and his orchestrated return to power, other significant GenAI players are moving forward decisively like the grown-ups in the room. The AI Alliance, recently announced by 50+ collaborating vendors, universities, and research institutions, including IBM, Meta, and Dell, offers insights into the dynamics, value, and promise of this approach.

Necessities of GenAI

Before delving into the AI Alliance, it is worth considering what constitutes and is required for developing successful and valuable GenAI platforms and services. In essence, GenAI requires access to:

1. Technical expertise/insights – Given the complexity of project design and development, it is arguable that no project can proceed, let alone succeed, without technically proficient and insightful teams and leadership. GenAI projects take years to develop and refine. In fact, the refinements never really end. So astute, dedicated leaders and teams are not simply desirable. They are necessities.

2. Computation – Similarly, computational resources are critical to GenAI project success, particularly if organizations hope to deliver commercially viable services and solutions. Without access to substantial computation, projects risk time-to-market delays that can translate into abject failures.

3. Training data – In terms of GenAI success, finding and developing data sets for large language models (LLMs) can be a double-edged sword. First, organizations can stumble during the process of searching for and assembling data sets. For example, OpenAI is facing numerous legal challenges from authors and organizations who claim the company is using its copyrighted intellectual property (IP) to train its ChatGPT chatbot and other projects. Additionally, poorly assembled or badly curated LLMs can result in AI hallucinations—statements from GenAI chatbots that are difficult or impossible to predict and demonstrably false or nonsensical.

4. Funding – Given the challenges and complexities inherent in the three previous points, access to substantial funding is a crucial requirement for Gen AI projects, especially those aiming to succeed in commercial markets.

These points aside, it is also worth noting the value that a fifth issue, openness, can provide GenAI and other artificial intelligence efforts. The past decade has witnessed the ongoing consolidation of technological resources and financial power among a handful of increasingly massive corporations, all of which are pursuing GenAI.

In essence, the transformational GenAI “revolution” that many in the tech industry extoll has devolved, for some, into a race to see who can get viable enterprise- and consumer-focused services to market first. In contrast, supporting and utilizing open GenAI technologies should help level the playing field and encourage the participation of and innovations from other commercial, educational, and research organizations.

The AI Alliance

What exactly is the AI Alliance? The group includes a wide range of AI-focused tooling creators, universities and science agencies, hardware, software, and silicon vendors, framework developers and promoters, and open model creators.

What goals and efforts are AI Alliance members, including IBM, Meta, and other members and collaborators planning? According to the announcement, the Alliance is aiming to foster “an open community and enabling developers and researchers to accelerate responsible innovation in AI while ensuring scientific rigor, trust, safety, security, diversity and economic competitiveness.”

Plus, the organization aims to “empower a broad spectrum of AI researchers, builders, and adopters with the information and tools needed to harness these advancements in ways that prioritize safety, diversity, economic opportunity and benefits to all.” Finally, members believe that “more collaboration and information sharing will help the community innovate faster and more inclusively, and (will help) identify specific risks and mitigate those risks before putting a product into the world.”

In order to achieve these goals, the AI Alliance will begin or enhance a wide range of projects, including benchmarks and evaluation standards, open foundation models, an AI hardware accelerator ecosystem, and global AI skills building, exploratory research, and educational content and resources.

Those are all admirable points. Plus, in sum, the companies, organizations, and groups involved in the AI Alliance possess the four key necessities for GenAI development, as well as a dedication to open technology sharing and collaboration.

Final analysis

What does this mean in real, practical terms? That might be considered by the companies and organizations pursuing substantial efforts around GenAI, which are notably absent from the AI Alliance. Those include Cisco, Google, HPE, Lenovo, Microsoft, NVIDIA, OpenAI, Salesforce and X.

It would be a mistake to read too much into the organizations missing from that list. However, Microsoft/OpenAI, Google, and X are all moving forward with GenAI projects and services that are essentially proprietary. Microsoft is already delivering products based on its $10B investment commitment to OpenAI. Those include the OpenAI Tools and developer/compute services the company offers on the Azure cloud platform, as well as the new ChatGPT-enabled Copilot solutions embedded in Office 365.

Google launched its Bard chatbot in February and has updated the service several times since then. Recently (in September), the company enabled Bard to support services, including YouTube, Google Drive, Google Flights, and others, and then (in October) launched Google Assistant with Bard. During the past week, X launched its Grok chatbot as a feature for Premium+ members of X (formerly Twitter).

These are interesting developments, but it is important to remember that the benefits of “first mover” status tend to be overly fetishized in the IT industry. Sure, vendors often enjoy significant benefits for achieving and delivering innovative new products and services. Apple’s iPhone is the poster child for turning technological innovations into notable commercial success.

However, more complex technologies and situations, including artificial intelligence, typically follow different and more measured paths. AI technologies have been explored and developed since the 1950s by a wide range of vendors, researchers and others. However, AI remained a sideroad in comparison to other, more commercially viable development areas.

While Microsoft, OpenAI, Google, and others deserve kudos for moving forward so quickly, the companies and their services hardly represent the ultimate endpoint of GenAI evolution. Plus, emphasizing speed over responsibility and trust (as the litigation OpenAI faces suggests) can place unnecessary barriers in the way of innovation.

The fact is that artificial intelligence, including offshoot technologies such as GenAI, is still in its relative infancy. How far and how fast those and other technologies will mature and evolve is anybody’s guess. IBM, Meta, and other members of the AI Alliance believe that careful collaboration and dedicated openness offer better, surer ways to pursue that journey.

Scroll to Top