OpenAIistakingonthevoiceassistantmarketwithitsnewVoiceEnginetoolandmayhaveApple,AmazonandGoogleinitssightsforthenextbigartificialintelligencepush.

ChatGPThasavoice-friendlyinterfaceonmobileandrecentlyintroducedawaytospeakresponsesondesktop,butOpenAI'snewtrademarkapplicationforthetermVoiceEngineisspecificallyrelatedtobuildingdigitalvoiceassistants.

Androidnowallowsuserstoreplacethedefaultvoiceassistant.AppleappearstobeindiscussionswithvariousAIcompaniesaboutthefutureofartificialintelligenceintheiPhone,andthismaybeapreemptivemovebyOpenAItobuildapotentialnewmarket.

AppleisalsorumoredtobeopeninganAI-onlyAppStorewiththenextmajorupgradeofiOS,whichwouldcreateanewmarketforAI-poweredassistants.

OpenAICEOSamAltmansaidthereare"avarietyofthings"thatwillbereleasedthisyear.ThisisexpectedtoincludeSora,anAIvideotool,butcouldalsoincludeanewAIvoicesystem.

Wedon'tknowmuchaboutVoiceEngineorevenifitwillbeaproduct;OpenAIhasnotcommentedpubliclyonit,soallwehavearerumorsandtrademarkapplications.

VoiceEnginemaybeanewmodelcreatedspecificallyforvoiceapplications,butitcouldalsobepartofOpenAI'senterpriseplay.Ithasthepotentialtobuildahigh-qualityvoicesystemthatwillallowcompaniestobuildmoreefficientcallcenterbots.

AnewtrademarkapplicationwasfiledlastweekwiththeU.S.PatentandTrademarkOffice.Whilethefilingdoesnotnecessarilyimplycommercialization,itisconsistentwiththebroadermarketshifttowardvoiceandthedirectionofOpenAI'stargetmodel.

Theapplicationcoversbuildingdigitalvoiceassistants,generatingspeechfromtextprompts,processingvoicecommands,andcreatingsoftwareusedtoprovidevoiceservices.

Thecompleteapplicationcoversthedevelopmentofvoiceserviceprovision,theuseofAIfortextorspeech,texttospeech,naturallanguage,andspeechprocessing,thegenerationofspeechandvoicefromprompts(text,audio,visual,images),voicecommandprocessing,speechrecognition,andbuildingdigitalvoiceassistants.Coverage.

Thissoundslikeallthepiecesneededforafullyfunctional,fullyinteractiveAIvoiceassistantthatcanhandlecomplextasks,chatnaturally,andeventakephonecallsforyou.

OpenAIreleasedGPT-4ayearago.Atthetime,thiswasthegroundbreakinggenerativeAImodelthatpoweredChatGPTandMicrosoftCopilot.

ThecompanyalsobegantrainingGPT-5latelastyear,sparkingspeculationaboutitsreleasedate.AltmantoldpodcasterRexFridmanthatthecompanywill"releaseanamazingnewmodelthisyear,"butdidnotconfirmwhetherthisistheGPT-5orapreliminarystep.

Healsosaidthat"variousthings"wouldbereleasedinthecomingmonths;accordingtoOpenAICTOMiraMurati,thisincludestheAIvideoplatformSora.

SomeonsocialmediahavespeculatedthatSoraandthisnewVoiceEnginearedifferentmodalinterfacesforGPT-5.

ItisverylikelythatGPT-5willbeatruemultimodalmodel,abletounderstandvideo,images,voice,text,andcode,aswellasgenerateallthesetypesofcontent.

Giventhetrademarkdescription,VoiceEnginecouldalsobethenewvoiceassistant,mergingthebroadcapabilitiesofSiri,Alexa,orGoogleAssistantwithChatGPT'sreasoningandnaturallanguagecapabilities.

GooglehasalreadybegunupgradingGeminitoworkthatway,AppleisrumoredtobebuildinganewversionofSiriwithextensivelanguagemodelingcapabilities,andAmazonisalreadytestingAlexaPluswithsimilarunderlyingskillsThetwocompanieshavebeenworkingonanewversionofSiriwithalargelanguagemodelingcapability.

OpenAImightofferaVoiceEnginetorunsuchasysteminthefuture,orasanalternativeinterfacetoChatGPTthatworkswithsmartspeakers,phones,orheadphones.

OritcouldjustbethatOpenAIisbeingcautiouswithitstrademarks:itsbidtoprotectGPTwasrejected,anditisnowapplyingfortrademarksGPT-5,6,andevenGPT-7.Thelatterincludesthegenerationofmusic,theconversionoftextanddataintocode,andthecreationofcodefromscratch.