Google is making changes to its AI search feature, AI Overviews because it has been giving some strange and incorrect answers to users. For example, it advised the public to eat rocks, it also recommended that people should stick glue on the pizza so that the cheese may stick better. This was literally launched at Google I/O 2024 where it was aimed to offer summaries of topics to be searched through the use of AI. At first launched with access to hundreds of millions in the United States, it was set ideally, to exceed a billion users in the coming year.
The company is trying to improve its search engine with AI as it faces competition from companies like Microsoft and OpenAI. Nevertheless, in the last week, there were screen captures of strange answers from AI Overviews recently posted by users.
There is a Google, blogpost by Liz Reid, the search head which was written on May 30, elaborating details that they have put mechanisms that help in avoiding nonsensical queries to depict an AI Overview. They have also made a decision not to permit what was regarded as humorous or satirical in the previous years.
Google has updated its systems to reduce the use of user-generated content that could lead to misleading advice and added restrictions for queries where AI Overviews were not proving useful. Reid pointed out that less than a handful of the total Overviews constructed using Artificial Intelligence violated content policies that transmit possibly toxic or pornographic content. Ultimately, she said that comparing the percentage of unique queries that break these policies to a chance, the rate is less than one in 7 million.
These changes come at a time when Google has set its sights on making money out of AI features for search and shopping in aspects such as AI Overviews providing tested and approved search and shopping ads. Reid blamed these inaccurate results on “data voids” or “information gaps,” where there’s limited high-quality content on a topic. For instance, the idea to eat rocks was acquired due to translated content that was initially reposted on a satirical resource related to geology, and the case with glue on pizza originated from sarcastic comments in online forums.
Reid went further pointing out that these Overviews do not ‘hallucinate’ which in essence means that the Overviews are not wolves that make things up. Such mistakes stem from misunderstandings between the real query and the implementation, between different languages, or between good and bad information. She compared AI Overviews to “featured snippets,” a similar search feature, claiming their accuracy is comparable.
Before releasing a feature called AI Overviews, Google successfully tested AI Overviews in terms of evaluations & tests, in addition to the analysis of search traffic. Here, Reid stated that additional and sometimes stupid usage by millions of people in the real world has forced new and nonsensical questions that seek to create wrong answers. Chen Lin also mentioned that some fake screenshots have appeared, and people thought they received dangerous advice from AI Overviews.
Though there are such problems Reid also said that the user feedback suggests that, people are more satisfied with the search results owing to the introduction of the AI overviews. Performance metrics are increasing, Teijiro Othera notes that people ask more sophisticated questions and use Overviews to go to more interested URLs. Google also does not want to display the AI Overviews for the items related to topics that are significant in forming and updating the news, as well as for the items that contain fundamentally vital and essential information – such as health-related search results, where the quality has been adapted since 2019.
At large, Google continues trying to polarity its AI search in response to users’ complaints regarding unexorable and wrong answers. The revelation of key details illustrates how the company’s goal to enhance the efficiency of AI-generated summaries for content that is relevant in searches can help prevent such situations by minimizing the number of articles with this type of content published and refining the mechanisms for detecting them.
Forget Buttons! Will Advanced AI Turn Siri into Your Ultimate App Assistant?
Apple plans to make massive upgrades to its one-billion-dollar digital invention, Siri, through the integration of artificial intelligence (AI). This update reported by Bloomberg will enable users to comprise certain functions of an app using voice instructions.
Siri’s AI Overhaul
Over the next year, Siri will be able to handle every control within an application. This enhancement forced the company to recode Siri’s program from scratch to incorporate large language models; technology that is behind generative AI. This update forms part of Apple’s Met Geraldine and was doubled as indicating that the company will incorporate even more AI functionalities into its devices during the Worldwide Developers Conference on the 10th of June.
New Features and Capabilities
Apple is preparing several new features for Siri, including:
Creating a voice memo often means creating a voice recording and enlists the use of the user’s voice to record brief notes or information which can be reviewed during leisure moments or even other times when information is unsuitable to take down on paper and write down while away from one’s blackberry.
This is due to their capability in enabling users to have quick overviews of websites and notifications.
Automating replies to messages
Advanced photo editing
Creating AI-generated emojis
Some of the basic Artificial Intelligence operations will reside on the devices themselves, while the advanced computations will be done through cloud computing. Shk is also negotiating with Open AI to have the chatbot and other services from the company to iOS. Meetings are under discussion with Alphabet Inc. over the possibilities of utilizing their Gemini system as well. Apple’s software frontman, Craig Federighi, spearheaded the development of many new AI features for this year’s operating system releases.
Enhanced Precision and Functionality
The new system will allow Siri to perform more precise tasks, such as:
-Opening specific documents
-Transferring music collections from one folder to another
-Sending or deleting emails
This feature involves opening specific articles in Apple News that the user has already set restrictions on by only giving access to the allowed domains.
Summarizing articles
These updates serve the goal of enhancing the capabilities and utilization of Siri to become a smarter and more sophisticated assistant with enhanced ease of use as opposed to simply controlling basic functions like playing music and adjusting the smart home gadgets and appliances. The AI features will first only be provided for native Mac apps, however, more commands are expected to be supported in the future.
The implications for privacy in future development
This topic leads to future plans for the development of relevant technologies and potential privacy issues.
These updates are likely to be released gradually by Apple, as articulated above. Apple is expected to roll out the first edition of the new operating system in September with the next iPhone variants. First of all, Siri is programmed to execute one command at a time, although latterly the ability to chain those commands will be available. For example, Siri could transcribe a recorded meeting and send a message containing the transcription to one’s assistant in one go.
The original adaptation will have AI that will determine whether the tasks should be done on the device or through the cloud. Here onsite tasks will not send any personal details while cloud-based tasks will have to send some of the data to remote servers. Apple has promised to defend this information using the Secure Enclave technology and is going to convince consumers by using the intelligence report concerning their pertinent security policy.
Revitalizing Siri
This is an attempt to put the ‘sparkle’ back in Siri, which has been lagging behind its innovative counterparts like Alexa and Google Home. Siri was launched in 2011 and it provided Apple with a head start specifically in the voice-based AI systems which Apple has been lacking behind in most of these rivals. In this case, the new AI features are aimed at compelling users to update their devices, because some of these assets will work only on the newest devices.
Conclusion
The new AI developments that Apple has planned in development for Siri will likely help this virtual assistant become a much more effective tool when used for helping people move about in iPhones and iPads. With the help of modern AI solutions and providing users with data safety, apple aims to recapture the leadership in the virtual assistant market and stimulate the sale of its equipment.
Can We Handle It? UN Sounds Alarm on Humanity’s Race to Harness AI
Speaking at the Abu Dhabi International Robotics Conference, the head of the United Nations International Telecommunication Union (ITU), Doreen Bogdan-Martin, said that the world is in a state of a contest to adopt the powers of artificial intelligence and use them for the benefit of society while excluding the possibility of its negative impacts.
The prospects of AI are getting ever brighter, and the question of its long-term future will continue to be debated for years to come.
Bogdan-Martin said this in her speech at the AI for Good Global Summit in Geneva stating that the recent developments in the field of AI are exceptional. However, in the same speech, she cautioned that reigning in the power of AI has become an emergent issue. Now there is no going back and one can only imagine how it is going to be, she said thus making her point that there is a need to act now.
Opportunities and Challenges
Climate change, extreme poverty, global food scarcity, elder care, and similar challenges have presented themselves as solvable problems through the use of AI. As the world is going through rapid transformative changes in every sphere of life, Bogdan-Martin feels this is a good time to make sure everybody gets to reap good values from AI. At the same time, she also pointed out that approximately 50% of the world’s population continues to be inactive and deprived of artificial intelligence, which she considers to be inconceivable.
Artificial intelligence has been said to be a double-edged sword in as much as it has got a lot of benefits it has also got devastating impacts that come with using it.
Bogdan-Martin in particular pointed out that although artificial intelligence technologies are double sided. AI should be created safe and secure as there is expected a very big year for elections in 2024 with many countries including the United States of America heading to the ballot. She questioned deep fakes and disinformation including but not limited to fake documentaries, the destruction of democracy, insecurity among the youth, and compromised cybersecurity.
Need for Regulation
The ITU chief observed that while AI is fast gaining influence in global affairs, the power to wield it remains in the hands of a select few. She also expressed some positive feelings about the recent activities of governments and organizations in placing checks and balances on AI affairs and users. For example, the European Union recently declared that the legislative initiative is working on developing an AI Office and that AI will be governed under a new universal law.
In conclusion, Bogdan-Martin summed up her discourse by urging people to collectively pen the next chapter of the tale where mankind and technology would be the protagonists. She also called for the advancement of AI in some ways that can be safe, inclusive, and sustainable so that the impact of this application can be felt by everyone and at the same time replicate the drawbacks of the application to the minimum level.
The current actions for AI should shape humanity to be all-inclusive and uplifting to enhance the course of the future.
Check who got the $$ Spotlight $$ today!
- HoundDog.ai Raises $3.1M in Seed Funding. read more
- Index Ventures backs AI manufacturing startup EthonAI in $16 million funding round. read more
- Plenful Secures $17M Series A to Streamline Pharmacy Operations with AI-Powered Workflow Automation. read more
- Zendata Scoops $2 Million Seed Funding to Address AI and Data Governance. read more
- Inspeq AI raises $1.1M to safeguard the future of AI with Sure Valley Ventures. read more
- Intelligent AI receives £800k funding boost. read more
- Sander Renes Receives NWO Funding for Automating Economic Behaviour Experiments with AI Agents. read more
- Weave Bio Announces $10M in New Funding and Launch of Its AI-Powered Platform to Streamline Drafting, Reviewing & Submitting Regulatory Documents in Drug Development. read more
- Female-led London’s Exactly.ai snaps $4.3M to empower artists with AI-powered design creation. read more
- Lithuanian Firm UDS Secures Funding to Enhance AI-Powered Swarm Drone Technologies. read more
- £250k funding for Liverpool elderly care AI companion Lola. read more
- Bain Capital Ventures in talks to lead Poolside AI’s $450M round. read more
- AI-powered precision agriculture platform Fyllo raises $4 Mn funding led by IndiaQuotient, SIDBI Ventures. read more
- ExpressionEdits Raises $13M Funding for Protein Expression. read more
- Ducky raises $2.7 million for AI customer support platform. read more
- Maxa Closes Series A Funding as Part of US$21 Million Round, and Brings Enterprise Data to Boardroom and Gen AI Conversations. read more
- Index Ventures Backs EthonAI in the Acceleration of AI Manufacturing Startup with $16 Million Funding. read more
- Business banking start-up Relay raises $32.2m in Bain Capital Ventures-led Series B. read more
