2025 : The trough of disappointment… or an opportunity for a new approach to AI?

2025 : The trough of disappointment… or an opportunity for a new approach to AI?


The Edelman barometer of trust report always makes for good reading.? The 2025 edition is no different.? And being involved in the AI field myself, I found the combination of these two results very interesting.?

For one - we trust businesss.

But we don't trust the AI solutions ... (produced mostly by by businesses).

I will leave it to others to speculate on what this means for the sky high valuations that are "AI washed" these days. My question though is - Why? And what can be done about it?

I think the reasons for "Why" are reasonably well known. Some key ones :

  • Lack of Transparency: Many AI systems, particularly deep learning models, function as "black boxes." This means it's incredibly difficult to understand how they arrive at their conclusions. This lack of transparency makes it hard to trust the system, debug errors, and ensure fairness and accountability. Without understanding the reasoning behind an AI's decisions, it's difficult to identify and correct biases or ensure that the system is acting ethically. Remember, atleast for now, all output is directed at human action / affecting real humans...if they don't understand it, they won't trust it.

  • Hallucinations: AI models can sometimes "hallucinate" – they can confidently produce incorrect or nonsensical outputs. The models can produce outputs that are incomplete, contradictory, or simply wrong. These hallucinations can have serious consequences, especially in critical fields like LifeSciences. Haven't we all had experience of a friend or a salesman making claims which are either immediately (or more dangerously after a period of time) discernable as falsehoods? How often will you go back to them for trusted advice after this happens once or twice??

  • Data Privacy and Security: AI systems often require vast amounts of data to function effectively. This data can include sensitive personal information, such as medical records, financial data, and even location history. This raises serious concerns about data privacy and security. There's always the risk of data breaches, where sensitive information could be accessed by unauthorized parties. Additionally, there are concerns about how this data is used and who has access to it. If were dealing with humans, the English language has a rich vocabulary of words to describe such people ;-)

What can be done about regaining and improving trust in AI solutions? Three things :

  1. Humility : Builders of AI solutions should stop pretending that the solution is all powerful and can solve world hunger, climate change & bring global peace - all within a day ;-) . We need to be transparent on the limits of our technology and work with customers to drive adoption
  2. People in the middle : "Man in the middle" is an old cybersecurity challenge. I think "people in the middle" is the solution to AI challenges. Build solutions for people to use...enable them, don't look to replace them. Take away the most tedious and boring and error prone elements of their daily lives and given them contextual assistance to help drive them to be more successful in their tasks
  3. Focus on value: Trust comes from making and keeping committments. In the business context, I translate that to understanding and delivering value. However "small" it may be, understand what's valuable to the client, be honest about what difference you can make and then deliver on that promise. Trust will follow.

What do you think? How can we collectively help improve adoption and build trust?

Pranjal Das

GTM Strategy & AI-Driven Sales Automation | 10+ Years in CRM, Demand Gen & Business Growth | Microsoft Dynamics 365, Salesforce, HubSpot | AI-Powered Insights & Data-Driven Decision Making

1 个月

The idea of "improving trust in AI" resonates deeply with me. Trust is the foundation for meaningful innovation, and as we approach 2025, fostering transparency and accountability in AI will be crucial. Excited to hear more perspectives on this.

Atul Gupta

Consultant, Architecting on DPI & DPG for National Scale, Certified Independent Director, Certified ESG Expert, Certified Digital Director, Cloud Architect, Leadership, Delivery, Technology & Innovation, Mentor

1 个月

Srinivas Padmanabharao a new technology is never without its challenges. Right now, to me AI is in the hype cycle. Everyone wants to have a skin in the game for the fear of being left out, but most don't either have anything concrete or are still grappling with what to make of it all. The issues and challenges with the technology are very real, but I do feel that the space is very powerful and it will evolve and see better solutions coming out soon. I am more worried about 1. what does it means for job market. Many jobs will be lost because where ever it was head count based, things will be done better and faster with AI assisted solutions and hence need for lesser workforce. More people become unemployed. How do they get their livelihood? People say new opportunities will come, but no one really has clarity. 2. where are we headed? I cannot keep myself from thinking of all the sci-fi movies where machines eventually take over. ??

要查看或添加评论,请登录

Srinivas Padmanabharao的更多文章

  • The "LLM" approach to AI in the real world

    The "LLM" approach to AI in the real world

    Last week I had written about the issue of trust in the AI. Over the weekend, I had a chance to read the US Government…

  • Happy New Year

    Happy New Year

    Yes, I know that it is only Nov 29th… but then have you forgotten that we will be entering the 3rd year of the AC era…

    3 条评论
  • Do you remember what you saw at SITC?

    Do you remember what you saw at SITC?

    Thanks to Society for Immunotherapy of Cancer (SITC) for a great #SITC2024 . There were over 1400 posters and equally…

  • Got a booth? get Knolens Edge !!!

    Got a booth? get Knolens Edge !!!

    As an exhibitor attending Society for Immunotherapy of Cancer (SITC) or sponsoring a booth at #sitc2024 in Houston this…

  • Welcome to SITC 2024, Houston

    Welcome to SITC 2024, Houston

    You have a camera on your mobile phone. Are you ready to use it to drive greater value from your investment of going to…

  • Revolutionizing Systematic Literature Reviews

    Revolutionizing Systematic Literature Reviews

    In the pharmaceutical industry, systematic literature reviews (SLRs) are indispensable for gathering and analyzing…

  • AI : The bottom line challenge

    AI : The bottom line challenge

    It's been nearly two years since ChatGPT burst onto the scene, ushering in a new era of artificial intelligence (AI)…

  • Lets beat infectious diseases !!!

    Lets beat infectious diseases !!!

    Let's wind back the clock. In 1978, CDC set a goal to eliminate measles from the United States by 1982.

    3 条评论
  • Instant Insights at your fingertips

    Instant Insights at your fingertips

    Instant Insights for the biomedical professional : AI at your fingertips Imagine you’re at an important industry…

  • Analytics in the age of LLMs

    Analytics in the age of LLMs

    Once upon a time, most of our communication used to be via spoken voice (the telephone, voicemail or face to face) and…

社区洞察

其他会员也浏览了