The major issue right now appears to be that costs of implementation far exceed expectations, in addition to looming legal and regulatory threats. The question now is at what point a sustainable path to profitability will emerge to justify the excitement. JL
Maju Abraham reports in LinkedIn:
Gen AI implementation is not without challenges, including getting users, both internal and external, to trust and accept AI-generated content. There may be skepticism and resistance to relying on AI for critical product development tasks. Other issues include data quality and quantity (ie, availability and cost); ethical and bias concerns; customization and fine-tuning requirements are resource-intensive; integration with existing workflows can be disruptive, requiring significant training and change management; lack of domain specific knowledge affects depth and context; intellectual property rights legality concerns; quality assurance and validation (can be time consuming and expensive); security and privacy violations; cost of implementation, including initial set-up and ongoing maintenance can be (daunting).
It’s hard to believe, but ChatGPT was launched over a year ago, in late November 2022 — thrusting generative AI into the mainstream. Since then, nearly every tech giant, from Microsoft to Google to Amazon, has taken a leap onto the Gen AI bullet train.
Still, as technologists discover more and more use cases for saving time and money in the enterprise, schools, and businesses the world over are struggling to find the technology’s rightful balance in the “real world.”
As the year has progressed, the rapid onset and proliferation has led to not only rapid innovation and competitive leapfrogging, but a continued wave of moral and ethical debates and has even led to early regulation and executive orders on the implementation of AI around the world as well as global alliances — like the recent Meta + IBM AI Alliance — to try and develop open frameworks and greater standards in the implementation of safe and economically sustainable AI.
Nevertheless, a transformative year with almost daily shifts in this exciting technology story. The following is a brief history of the year in generative AI, and what it means for us moving forward.
How to accelerate a successful implementation
Gen AI could be the most disruptive technological innovation in the next 5 years.
- Understand possibilities – superior outputs faster
- Extract value – Quick wins, long term
- Every day tasks
- Reshape critical functions
- Invent new business models
- Deploy what is right for your business
- Modernize your tech stack for scaling generative AI
- Invest in people and process
“One of the most overlooked part of an AI implementation is adoption. Organizations are effective in deploying AI solutions but fail to achieve the intended value due to lack of adoption by the workforce.”
The Year In Gen AI: How It Started
As a brief reminder, generative AI, a.k.a. “Gen AI,” took the world by storm a year ago with the advent of ChatGPT, a technology launched by artificial intelligence developer OpenAI. Within days, ChatGPT changed the way we think about everything, from business to education, allowing the average person to write almost anything – letters, memoirs, TV scripts – with the click of a button.
All a user needs to do is input a few key prompts (i.e. “Create a movie script similar to Alien with a younger cast, set in 2023 …”) and ChatGPT spits it out, almost instantly.
The technology itself was mind-blowing and almost immediately promised to displace everything we as humans have come to take for granted — from journalists to Hollywood writers to the standard college essay. After all, to the common eye, there’s almost no way to determine if something was created by AI or written by a human being. The issue set off not just alarm bells regarding the moral use of generative AI, but a nearly five-month walk-out by TV and movie writers concerned that AI could make their own use cases extinct.
The thing is, as we’ve learned in the past year, generative AI is so much more than ChatGPT. It encompasses everything from apps that record and transcribe conversations in meetings to apps that draft email replies, create voiceovers, and generate “original” art and logos. Generative AI has the power to create just about anything, offering the potential to save time and money for today’s businesses, while also creating new opportunities for growth and investment.
In fact, since the release of ChatGPT in November 2022, research found that Gen AI features could add up to $4.4 trillion to the global economy annually. No wonder it’s been getting so much attention–for better or worse.
The Year In Gen AI: How It’s Going
Just a few months after the launch of OpenAI’s ChatGPT, Microsoft announced that its Bing search engine would be adding a ChatGPT element. While the declaration didn’t boot Google from the top of the search engine list, it did initiate a series of new announcements from other tech giants rushing to join the gen AI game. In fact, the Bing/ChatGPT partnership launched a seemingly endless wave of Gen AI developments over a series of just a few months. Here are a select few:
Microsoft
Microsoft announced a slew of additional OpenAI-powered elements, including the addition of a new Copilot feature for both Microsoft Dynamics 365 and Microsoft 365.
The Copilot feature helps business and consumer users do things like summarize and edit text, create a more casual tone for an emails, gain insights from data sets, and even create presentations based on content created in other Microsoft apps like Word and Excel. Microsoft added capabilities across its portfolio at a torrid pace.
Google, while getting off to a slow start with a somewhat awkward Bard launch was ultimately not to be outdone. The company began offering Duet, a technology similar to Copilot, that allows users to harness the power of Gen AI throughout the Google Workspace, helping with writing and even creating custom visuals in presentations, etc.
Google also launched Generative AI Studio, which allows developers to create their own generative AI apps with text and images, even if they aren’t super familiar with AI or machine learning. This past week Google launched Gemini with an impressive demo that has brought excitement to the Gen AI race, while also raising some questions about how much of the demo is really achievable today.
Salesforce
Salesforce announced Salesforce AI Cloud, which allows for a more tailored sales CRM experience where commerce teams can request auto-generated insights on improving user experience on the customer journey; create personalized content for email, web, and mobile; create replies for customer service teams; and generate personalized emails for customers. The company has built its story around safety of generated data on its platform.
Adobe
Adobe announced that its Adobe Firefly would allow users to create images, 3D imagery, vectors, and even audio and video content with a click. While language models have been the focus of much of 2023, image based generative tools have big implications and Adobe is well situated to be the leader in this space.
Amazon Web Services
AWS joined the GenAI conversation this year but made its big splash in late 2023 at its reInvent conference announcing not only its own Titan model, but also its open model approach as well as its new Q offering, which is perhaps most simply put, its LLM for enterprise. The company also launched its next generation training chip, Trainium2.
IBM
IBM announced the development of new platforms like its multi-faceted watsonx to support more powerful AI base models. The company also started a wave of indemnification offerings for the use of GenAI tools as the speed of innovation created legal concerns for security, data rights, and potential infringement — being first to market this capability as it was also first to GA its enterprise offering watsonx.AI.
Nvidia
Nvidia dominated the silicon for GenAI, seeing an unprecedented rise in demand for its most advanced GPUs while its competitors raced to catch up. AMD just launched its newest GPU to offer Nvidia some competition while more efficient, less flexible chips from AWS, Google, Microsoft, and Intel also saw market adoption.
We also saw a rise in attention toward silicon for data transport and infrastructure with companies like Broadcom and Marvell seeing a substantial uptick in attention. To be clear, silicon is at the core of all of these GenAI tools and innovation, supply, and competition is very important here.
In other words: by early 2023, the tech world itself was virtually synonymous with Gen AI. Never in history has another technology had this much impact, this quickly. And while that list had a lot of generative AI updates, it was far from exhaustive and barely scratched the surface on exciting companies that have an AI story to tell.
What The Data Is Saying About What’s Next For Gen AI
Over the past year, we have been tracking closely how AI and generative AI will proliferate in the enterprise. In November, our team at Futurum Intelligence published its first-ever AI Decision Maker in The Enterprise Study. Surveying over 1,000 AI decision makers in the enterprise, we ranked 159 different vendors of AI deployments, and the findings are astounding and very telling. Here are a couple of key takeaways.
1. Across the spectrum of companies, we saw a 300%+ increase in companies planning to make multimillion dollar investments into GenAI in 2024 versus 2023
2. The top 2 criteria for choosing vendors for GenAI projects were perceived expertise with AI and implementation speed and timeline.
While the findings are extensive, the TLDR here is that companies are ramping up investment at an incredibly fast rate, while pushing for faster implementation and time to value — both serving as positives for tech spend in 2024 and beyond matching the expected mid-double digit CAGR that have been reported by multiple sources for GenAI over the next five years.
The Year in Gen AI: Beyond Commercialization, Where Do We Go From Here?
In the past year, we’ve seen Gen AI grow further and faster than technologists likely ever anticipated. This begs the question: where do we go from here?
Just last year, OpenAI chief exec Sam Altman was ousted by the company’s board, only to be reinstated less than a week later. The ouster hints at the overall schism in AI’s development so far: the line that divides AI’s potential to help humanity and its potential to end it altogether.
Something we can all agree on: Gen AI isn’t going anywhere.
Even as issues associated with “unsupervised AI” continue to surface, the technology is advancing far more quickly than the human ability to place limits upon it. That means we’ll likely see Gen AI become even more accessible to the general public in the coming year. And, like we saw with last week’s passage of the AI Act in the E.U., we will hear more from leaders on regulation and compliance given the pace of innovation versus the risk to consumers.
Overall, it’s almost impossible to keep up with the rapid, almost daily meaningful advancements in generative AI. As AMD CEO Lisa Su alluded to at the company’s Advancing AI event last week, this is the fastest pace of change that the industry has ever witnessed—and it is safe to say, it will only continue to speed up from here. Exciting times ahead!
The Challenges of Implementing Generative AI in Product Development
Generative AI, with its ability to generate content, code, and even creative ideas, has been a game-changer in the world of product development. However, its implementation is not without its challenges.
1. Data Quality and Quantity:
Generative AI models require vast amounts of data to function effectively. Many organizations struggle with acquiring and curating high-quality data that represents their specific product domain. Without this foundational dataset, the AI model's outputs may lack relevance and accuracy.
2. Ethical and Bias Concerns:
One of the significant challenges associated with generative AI is the potential for bias in the generated content. If the training data used to build the AI model is biased, the AI can inadvertently produce biased or discriminatory outputs. Mitigating this risk and ensuring ethical use of AI is a complex challenge.
3. Customization and Fine-Tuning:
Off-the-shelf generative AI models might not be well-suited to an organization's specific needs. Customization and fine-tuning of these models can be a resource-intensive process, requiring expertise in machine learning and data science.
4. Integration with Existing Workflows:
Integrating generative AI into existing product development workflows can be challenging. Teams may need to adapt to new tools and processes, which can disrupt established practices and require significant training and change management efforts.
5. Lack of Domain-Specific Knowledge:
Generative AI models often lack domain-specific knowledge. While they can generate content, they may not understand the nuances or specific requirements of a particular industry or product category. This can lead to AI-generated content that lacks depth and context.
6. Intellectual Property Concerns:
When AI generates creative works or code, questions arise regarding intellectual property rights. Determining ownership and licensing for AI-generated content can be legally complex and is an evolving area of law.
7. Quality Assurance and Validation:
AI-generated content needs rigorous quality assurance and validation processes to ensure accuracy and consistency. Ensuring that the AI is generating content that meets the organization's quality standards is an ongoing challenge.
8. Cost of Implementation:
Implementing generative AI requires an investment in hardware, software, and expertise. The initial setup costs and ongoing maintenance can be significant, and organizations must weigh these costs against the potential benefits.
9. Security and Privacy Concerns:
AI-generated content can sometimes inadvertently leak sensitive information or data. Organizations must ensure that AI systems do not compromise security or privacy standards.
10. User Acceptance:
Getting users, both internal and external, to trust and accept AI-generated content can be challenging. There may be skepticism and resistance to relying on AI for critical product development tasks.
Generative AI has the potential to revolutionize product development by automating tasks, enhancing creativity, and accelerating processes. However, organizations must navigate these challenges to harness the technology's full potential.
Overcoming these hurdles requires a multidisciplinary approach, involving data scientists, domain experts, legal professionals, and ethical AI specialists. As the technology continues to evolve, addressing these challenges will be essential for successful integration and realizing the benefits of generative AI in product development.
0 comments:
Post a Comment